Analysis Of Contributing Factors And Prediction Model For Retinopathy Of Prematurity

Objective: To explore the influencing factors of retinopathy of prematurity and construct a risk prediction model to identify preterm infants who present risk factors of ROP as early as possible,to apply effective treatment to them in a timely manner and avoid the appearance of serious sequelae.Methods: clinical data of neonates hospitalized at the NICU of the First Affiliated Hospital of Nanchang University from January 2013 to June 2022 were retrieved,and all children who underwent fundus screening were divided into ROP group(41 cases)and non ROP group(82 cases)according to the 2014 version of the screening guideline for ROP issued by the Chinese Medical Association.The basic information,postnatal treatment,and perinatal conditions of the pregnant mother of the two groups were collected,and the risk factors affecting ROP were explored by univariate analysis.Combined with clinic to replace univariate variables with P < 0.2 into lasso regression to remove more commonalities variables,the effective variables after screening were used to establish a risk prediction model of ROP by binary logistic regression analysis,and a linear plot was drawn to visualize it,and the discrimination and calibration of the model were evaluated.Result:1.A total of 123 preterm infants were included in this study.Ultimately 41 children with ROP were included in the study based on fundus screening results.The proportion of ROP in preterm infants with GA < 28 weeks was 80.00%(8 / 10),which was higher than that in preterm infants with GA between 28 and 30 weeks and between 30 and 32 weeks(all P < 0.05).The proportion of ROP among preterm infants with BW < 1000 g was 66.67%(8 / 12),which was significantly higher than that among preterm infants with BW between 1000-1500 g and > 1500 g(all P <0.01).2.As shown in univariate analysis,(1)the basic condition of preterm infants:GA,BW,Apgar score at one and five minutes;(2)Primary disease and complications in preterm infants: severe neonatal asphyxia,BPD,neonatal pulmonary hemorrhage,AOP,HIE,and weight gain 2 weeks after birth;(3)Interventions for preterm infants: duration of invasive and noninvasive mechanical ventilation,total oxygen inhalation time,maximum oxygen concentration,oxygen concentration ≥40% of the time,number of blood transfusions,total blood transfusions,and total parenteral nutrition time were statistically different between the two groups(all P <0.05).3.The binary logistic regression results suggested that small for gestational age(OR = 0.630,95% CI: 0.412-0.962),low five minute Apgar score(OR = 0.446,95%CI: 0.265-0.749),and long total oxygen inhalation time(OR = 1.024,95% CI: 1.002-1.047)were independent risk factors for ROP.4.The predictive equation established by logistic regression was Z = 19.174-0.463 x gestational age at birth-0.808 x 5-minute Apgar score + 0.024 x total oxygen inhalation time;the C-index of this predictive model = 0.81 > 0.75(95% CI: 0.729 to0.897);the calibrated C-index = 0.80 > 0.75(95% CI: 0.717～0.883).Plotting the calibration curve showed that the Apparent line overlapped with the Bias-corrected and Ideal lines.Conclusions:(1)The smaller the gestational age and the lower the birthweight,the higher the incidence of ROP.(2)Several factors are associated with the occurrence and development of ROP,and it is beneficial to delay premature delivery as much as possible,to shorten the duration of mechanical ventilation,to avoid prolonged oxygen use,to reduce the use of highly concentrated oxygen,to strictly control the blood transfusion pointer,and to start enteral nutrition as early as possible to reduce the occurrence of ROP.(3)Small for gestational age at birth,low Apgar score at 5 min,and long total oxygen inhalation were independent risk factors for ROP.The risk prediction model for ROP developed based on appeal 3 independent risk factors had a good discrimination and calibration,which can predict the occurrence of ROP before the1 st fundus screening is performed.More targeted screening of children with high-risk factors for ROP may provide support for further individualized diagnosis and treatment.